distilled-mt5-small-b0.04 / trainer_state.json
Lvxue's picture
End of training
2f63853
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 5.0,
"global_step": 12500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.2,
"learning_rate": 4.8e-05,
"loss": 15.3489,
"step": 500
},
{
"epoch": 0.4,
"learning_rate": 4.600000000000001e-05,
"loss": 6.2024,
"step": 1000
},
{
"epoch": 0.6,
"learning_rate": 4.4000000000000006e-05,
"loss": 4.8868,
"step": 1500
},
{
"epoch": 0.8,
"learning_rate": 4.2e-05,
"loss": 4.2285,
"step": 2000
},
{
"epoch": 1.0,
"learning_rate": 4e-05,
"loss": 3.9284,
"step": 2500
},
{
"epoch": 1.2,
"learning_rate": 3.8e-05,
"loss": 3.6399,
"step": 3000
},
{
"epoch": 1.4,
"learning_rate": 3.6e-05,
"loss": 3.5056,
"step": 3500
},
{
"epoch": 1.6,
"learning_rate": 3.4000000000000007e-05,
"loss": 3.3825,
"step": 4000
},
{
"epoch": 1.8,
"learning_rate": 3.2000000000000005e-05,
"loss": 3.2741,
"step": 4500
},
{
"epoch": 2.0,
"learning_rate": 3e-05,
"loss": 3.1813,
"step": 5000
},
{
"epoch": 2.2,
"learning_rate": 2.8000000000000003e-05,
"loss": 3.1296,
"step": 5500
},
{
"epoch": 2.4,
"learning_rate": 2.6000000000000002e-05,
"loss": 3.0029,
"step": 6000
},
{
"epoch": 2.6,
"learning_rate": 2.4e-05,
"loss": 2.9858,
"step": 6500
},
{
"epoch": 2.8,
"learning_rate": 2.2000000000000003e-05,
"loss": 2.9246,
"step": 7000
},
{
"epoch": 3.0,
"learning_rate": 2e-05,
"loss": 2.9296,
"step": 7500
},
{
"epoch": 3.2,
"learning_rate": 1.8e-05,
"loss": 2.8465,
"step": 8000
},
{
"epoch": 3.4,
"learning_rate": 1.6000000000000003e-05,
"loss": 2.8507,
"step": 8500
},
{
"epoch": 3.6,
"learning_rate": 1.4000000000000001e-05,
"loss": 2.7587,
"step": 9000
},
{
"epoch": 3.8,
"learning_rate": 1.2e-05,
"loss": 2.8008,
"step": 9500
},
{
"epoch": 4.0,
"learning_rate": 1e-05,
"loss": 2.7439,
"step": 10000
},
{
"epoch": 4.2,
"learning_rate": 8.000000000000001e-06,
"loss": 2.7015,
"step": 10500
},
{
"epoch": 4.4,
"learning_rate": 6e-06,
"loss": 2.671,
"step": 11000
},
{
"epoch": 4.6,
"learning_rate": 4.000000000000001e-06,
"loss": 2.7106,
"step": 11500
},
{
"epoch": 4.8,
"learning_rate": 2.0000000000000003e-06,
"loss": 2.7473,
"step": 12000
},
{
"epoch": 5.0,
"learning_rate": 0.0,
"loss": 2.6899,
"step": 12500
},
{
"epoch": 5.0,
"step": 12500,
"total_flos": 2522355797975040.0,
"train_loss": 3.76287099609375,
"train_runtime": 2865.4475,
"train_samples_per_second": 17.449,
"train_steps_per_second": 4.362
}
],
"max_steps": 12500,
"num_train_epochs": 5,
"total_flos": 2522355797975040.0,
"trial_name": null,
"trial_params": null
}